Yes I read the manual. I'm not saying you're wrong, the board will work as you say, but the output buffers can only source 6mA at 3.8v typical output on a 5v supply. With a typical optoisolator circuit as in the MH drives (given a 390ohm input resistor and a 2v forward voltage) you'll actually get around 4.5mA , less than half what the optoisolator needs to meet the recommended drive levels in the MH driver spec, plus the active rising edge is slow due to the input capacitance which could become problematic at higher step rates. By sinking current you only lose 0.4v of the 5v in the buffer so you get closer to 7mA drive and better shape to the active falling edge so better at higher step rates and more resistant to noise on supply rails.

I guess the OP did wire it wrong as he took the + common to N and not to 5v output? If so I stand corrected on that point.